informa
/
2 min read
Feature

Supercomputers As Soothsayers

Researchers are making strides in using supercomputers to test materials and improve the ability to predict how new materials will withstand stress.
Researchers are making strides in using supercomputers to test materials and improve the ability to predict how new materials will withstand stress.

A team of scientists from IBM Research and Lawrence Livermore National Laboratory have released two research papers outlining their use of IBM's ASCI White supercomputer as a computational microscope to simulate the properties of materials, and calculate the interactions and behavior of as many as one billion atoms of a solid material. One test measured the travel rate of a brittle fracture, another examined ductile deformation, bending material without breaking it.

The supercomputer is able to calculate the force between the atoms and, using visualization software to simulate the behavior of atoms on a micron level, scientists were able to see brittle-fracture cracks as they traveled faster than previously theorized. The researchers now have evidence that a crack tip can shoot through the material faster than the speed of sound, something previously thought impossible. This result will help scientists understand a wide range of fractures, from shallow earthquakes to the sudden failure of fiber-reinforced composite materials that can be used in fields ranging from aerospace to chip design to medical devices.

Through simulating and visualizing how atoms move under stress, the scientists say the right amount of ductile deformation strengthens a material when done properly. They simulated a work hardening-the process by which deformation strengthens a material but can "embrittle" it if overdone-involving 1 billion atoms. Bending a paper clip, for example, demonstrates that the metal is initially flexible when bent but soon stiffens and breaks where it was repeatedly stressed, says Farid Abraham, a computational physicist at IBM Research. Previously, such testing was based largely on engineering principles for materials and the use of assumptions and trial and error for new developments.

The results are a major step to using supercomputers to design new materials with customized properties. A new alloy should be stronger, tougher, or lighter, Abraham says. "Now you can design the material and simulate properties on the computer."